February 22, 2015

The effects of aerobic high-intensity interval training on endurance performance. A systematic review of randomized controlled trials.

Michael Rosenblat

PLAIN LANGUAGE SUMMARY

High-intensity interval training (HIIT) involves performing a set of multiple bouts of exercise that are completed above the anaerobic threshold (LT2). HIIT programs can be differentiated by the dominant system used to produce energy during the interval session. Anaerobic high-intensity interval training (AnHIIT) involves exercise bouts that are less than two minutes in duration, whereas aerobic high-intensity interval training (AeHIIT) involves exercise bouts that are greater than two minutes. The purpose of this review is to systematically evaluate the available evidence on the effects of different AeHIIT training programs on markers of endurance performance in trained and highly-trained endurance athletes.
There is high level evidence demonstrating that AeHIIT can improve VO2peak in state/provincial-level male and female rowers.
There is moderate level evidence that demonstrates that AeHIIT can improve stroke volume in endurance-trained male athletes.
There is moderate level evidence that demonstrates that AeHIIT can improve LT2 (as a percent of VO2peak) in well-trained, competitive endurance athletes (middle distance runners, triathletes and 10 km runners).
There is both high and moderate level evidence that shows that AeHIIT can improve TT performance in well-trained, state/provincial-level male and female rowers, and endurance-trained male athletes, respectively.
There is moderate level evidence that demonstrates that AeHIIT can improve exercise efficiency in trained male and female cyclists, and endurance-trained male athletes.
There is moderate level evidence that indicates that AeHIIT improves TTE in trained male and female cyclists, and endurance-trained male athletes.
More high quality research is necessary to accurately assess the effects of AeHIIT on endurance performance. Specifically, this can be accomplished by designing studies that include blinded subjects and assessors, between-group comparisons, and intention-to-treat analysis. Finally, studies that include two or more exercise programs should use a method to calculate the training load.

INTRODUCTION

High-intensity interval training (HIIT) involves performing a set of multiple bouts of exercise above the anaerobic threshold (LT2). Each exercise bout is followed by a rest period consisting of either active or passive recovery. There are numerous factors to consider when designing a HIIT program. Buchheit et al 2013 explains that there are several variables that can be manipulated when programming for HIIT. These variables include interval duration, intensity, number of repetitions, duration of recovery period, intensity of recovery period and weekly distribution of interval sessions (1).
The goal of HIIT is to improve markers of endurance performance. Recent research suggests that HIIT can improve endurance performance beyond the level that results from traditional continuous low-intensity training. Furthermore, HIIT can accomplish this in a significantly shorter time (3, 4, 6).
HIIT programs can be differentiated by the dominant system used to produce energy during the interval session. Anaerobic high-intensity intervals training (AnHIIT) involves exercise bouts that are less than two minutes in duration whereas aerobic high-intensity interval training (AeHIIT) involves exercise bouts that are greater than two minutes. Both types of intervals are commonly described in the literature as a means to improve endurance performance. However, research that compares the mechanisms for which the two forms of intervals elicits physiological and performance adaptations is limited.
There are a number of reviews that describe the effects of interval training on endurance performance (1, 2, 3, 4, 7, 9). Currently, there are no systematic reviews of randomized controlled trials that specifically analyze the effects of AeHIIT on endurance performance. The purpose of this review is to systematically evaluate the available evidence on the effects of different AeHIIT programs on markers of endurance performance in trained and highly-trained endurance athletes.

Outcome Measures
The outcome measures included in the studies related to performance included: time-trial (TT), exercise efficiency and time-to-exhaustion (TTE). The physiological variables assessed in the studies included: relative VO2peak, LT2 (as a percent of VO2peak), and stroke volume.

Objectives
1.     Does AeHIIT have a greater affect on physiological markers of performance than other forms of endurance training in endurance athletes?
2.     Does AeHIIT have a greater affect on endurance performance in endurance athletes than other forms of endurance training?
3.     Is there a specific AeHIIT program (intensity/duration/load) that is optimal for improving performance in endurance athletes?

METHODS

Search Strategy
Pubmed and Medline were the databases used to perform the literature search. The search terms included: aerobic interval training or high intensity interval training. Limitations included: humans, English, randomized controlled trials.

Selection Criteria
Articles were selected if they included intervals greater than or equal to 2 minutes in duration, subjects between the ages of 18-65, trained or highly trained endurance athletes (VO2peak 50 ml/kg/min). Articles were excluded if they included sports that are considered anaerobic in nature, included untrained individuals or subjects with pathology.

RESULTS

Search
The literature search was conducted on January 31, 2015. Pubmed and Medline yielded a total of 283 and 90 articles respectively. Five articles were selected following the removal of articles that did not meet the selection criteria.

Study Quality
The PEDro scale (score out of 10) was used to rate the quality of the articles included in this review. The category “inclusion criteria” is not included in the total score. One study had a score of 7/10, two studies scored 5/10, and two studies scored 4/10 (Table 1).


Study Characteristics
All five of the studies selected analyze the effects of AeHIIT on VO2peak (Table 2). Three studies assessed the change in TT performance (2000m row, 3000m run and 40km cycle) (Table 2). Two studies also assessed TTE (running time at VO2peak and cycling time at 80% power output at VO2peak (PO@VO2peak)). Three studies analyzed the effects of different HIIT programs on efficiency (cycling in watts/kg, running in HR at 7km/h and running velocity at VO2peak). Two studies analyzed the effects of AeHIIT on LT2 (as a percent of VO2peak). One study compared the effects of training in different zones on stroke volume (ml/beat). The training protocols are described in detail in Table 2.




DISCUSSION

Does AeHIIT have a greater affect on physiological markers than other forms of endurance training in endurance athletes?
All five of the studies analyzed the effects of AeHIIT in VO2peak (5, 8, 10, 11, 12). The studies by Helgerud et al and Smith et al also assessed the effects of AeHIIT on LT2 as a percent of VO2peak (11).
            The study by Seiler et al evaluated three interval-training protocols (4-minute, 8-minute, 16-minute intervals) on 35 trained (mean VO2peak ~ 51.6 ml/kg/min) male and female cyclists over a 7-week period (10). Only the 4x16min group and the 4x8min group demonstrated a significant increase in VO2peak (+6.5%, p < 0.05 and +10.4%, p < 0.05, respectively) (10). Two of the subjects requested to switch groups due to time conflicts, introducing a significant bias with group allocation. Also, total training load was not kept constant among the groups. This makes it impossible to determine if the changes in VO2peak that occurred were a result of the interval duration or of the exercise. Finally, between-group comparisons were not conducted. Therefore, it is not possible to determine if one interval program is better than another for the improvements seen in VO2peak.
            Driller et al compared four weeks of AeHIIT with a continuous training program on ten, well-trained male and female state/provincial-level rowers. Training load was kept constant by using a TRIMP score (TRIMP = mean power x duration). The AeHIIT group completed 8x 2.5-minute intervals at 90% of peak power output (PPO). The control group completed approximately 40-50 minutes of low-intensity training with blood lactate levels between 2-3 mmol/L. The AeHIIT group demonstrated a significantly greater improvement in VO2peak (+7.0 ± 6.4%, p = 0.02) when compared to the control (5).
            Fifty-five male, endurance-trained (VO2peak mean ~57.7 ml/kg/min) university students participated in the 8-week trial by Helgerud et al. Training load was kept constant between groups by calculating the total oxygen consumption for each training session. The subjects were divided into four groups: 24 minutes at LT2; 15x 15 seconds at 90-95% HRmax; 4x 4 minutes at 90-95% HRmax; and a control group which completed 45 minute sessions at 70% HRmax. Only the 15x15sec group and the 4x4min group demonstrated a significant improvement in VO2peak (+5.5%, p < 0.01 and +7.2%, p < 0.001, respectively) (8). VO2peak in both the 15x15sec group and the 4x4min group were significantly (p < 0.001) greater than the LT2 group and the control group following the intervention (8). Both 15x15sec group and the 4x4min group also demonstrated a significant improvement in stroke volume (+9.4%, p < 0.01 and +10.4%, p < 0.001 respectively) (8). None of the groups showed an improvement in LT2 (8). Only 40 of the 55 subjects completed the study. The lack of adequate follow-up (80%) produces a significant issue with the quality of the results as the power calculation conducted to determine statistical significance was based on a sample size of 55 subjects. Also, an intention-to-treat analysis was not included. Therefore, by excluding subjects from the follow-up analysis, the main rationale for randomization was lost, increasing the risk of bias.
            Smith et al analyzed changes in VO2peak and LT2 following AeHIIT on 27 well-trained (mean VO2peak = 61.4(1.0) ml/kg/min), competitive athletes (middle distance runners, triathletes and 10 km runners). They conducted a 4-week trial, dividing subjects into one of three groups. The Tmax60 group completed six intervals at 60% of running time (~2:13) maintained at VO2peak; the Tmax70 group completed five intervals at 70% of time (~2:34) maintained at VO2peak; and the control group continued their current low-intensity training program. Only the Tmax60 group demonstrated a significant change (+6.8%, p < 0.05) in LT2 (11). There was no change in VO2peak in any of the groups (11). Between-group comparisons were not conducted making it impossible to determine if the intervals performed in the Tmax60 group are better than those performed in the Tmax70 group for improving LT2.
            Stepto et al included twenty provincial-level, endurance-trained (mean VO2peak ~61 ml/kg/min) male cyclists in a 3-week trial. The subjects were divided into five groups : 12x 30-second intervals at 175% PO@VO2peak followed by a 4.5-minute recovery period; 12x 60-second intervals at 100% PO@VO2peak with a 4-minute recovery period; 12x 2-minutes at 90% PO@VO2peak with 3-minute recovery periods; 8x 4-minutes at 85% PO@VO2peak followed by a 1.5-minute recovery; and 4x 8-minutes at 80% PO@VO2peak with a 1-minute recovery period. The results of the study showed no significant difference in VO2peak in any of the groups at follow-up (12).

Does AeHIIT have a greater affect on endurance performance in endurance athletes than other forms of endurance training?
Three of the five studies assessed changes in TT performance (2000m row, 3000m run, 40km cycle) (5, 11, 12). Two studies also assessed TTE (cycling time at 80% PO@VO2peak, running time at VO2peak) (10, 11). Two of the studies analyzed the effects of different AeHIIT programs on exercise efficiency (cycling in watts/kg, running in HR at 7km/h, 5.3% incline) (8, 10).
            The study by Driller et al demonstrated a significant difference in 2000m-row TT time between the AeHIIT group and the control group in favour of AeHIIT (-1.9% ± 0.9%, p = 0.03) (5). The results of the Smith et al study showed a significant decrease in 3000m-running TT time in the Tmax60 group (-17.6s (3.5), p < 0.05) with no significant change in the Tmax70 group (11). Stepto et al demonstrated a significant trend in improvement in 40km cycling TT in both the 12x30sec group (-2.8%, 95% CI = 4.3-1.3) and the 8x4min group (-2.4%, 95% CI = 4.0-0.7) (12). As previously mentioned, there were no between-group comparisons completed in either the Smith et al or the Stepto et al studies. Therefore it is unclear if the interval programs that demonstrated a significant difference at follow-up are truly better than the other programs included in their respective studies.
            There was a significant improvement in TTE in the study by Seiler et al in all three of the intervention groups; G4x16 (+62.3%, p < 0.05), G4x8 (+91.1%, p < 0.05), G4x4 (+63.3%, p < 0.05) (10). However, there was no between-group comparison. Smith et al demonstrated that six intervals at 60% of time maintained at VO2peak also improved TTE (+22.8%, p < 0.05) (11). Again, there was no between-group comparisons made in this study.
            Of the three studies that assessed changes in efficiency, only the studies by Seiler et al and Helgerud et al demonstrated significant changes at follow-up. All of the intervention groups in the Seiler et al study improved in cycling efficiency (watts/kg); G4x16 (+7.0%, p < 0.05), G4x8 (+12%, p < 0.05), G4x4 (+7.0%, p < 0.05) (10). All groups in the Helgerud study, including the control group, demonstrated a significant improvement in running efficiency (HR at 7km/h at 5.3% grade); LSD (-10 bpm, p < 0.01), LT2 (-14 bpm, p < 0.001), 15/15 (-22 bpm, p < 0.001), 4x4 (-18 bpm, p < 0.001) (8). There were no between-group differences in efficiency in either of the studies.

Is there a specific AeHIIT program (intensity/duration/load) that is optimal for improving  performance in endurance athletes?
Only two of the studies (Driller et al and Helgerud et al) included a method for monitoring training load. The study by Driller et al demonstrated that 8x 2.5-minute intervals at 90% PPO lead to a significant improvement in both VO2peak and in 2000m row TT time (5). Unfortunately, they only compared the HIIT program with a control group. The study by Helgerud et al compared both AeHIIT (4x 4-minute intervals) and AnHIIT (15x 15-second intervals) with a control group, as well as with group training at LT2. Both the AeHIIT and AnHIIT groups demonstrated a significant improvement in VO2peak when compared to the other groups (8). However, there was no difference between the AeHIIT group and AnHIIT group at follow-up.

Limitations
The current review systematically analyzed randomized controlled trials on the effects of AeHIIT on endurance performance. One limitation to the study design is that only two databases (Medline, PubMed) were used to conduct the literature search. This may limit the number of studies included in the review and increases the risk of selection bias. The main tool used to critically appraise the articles included in the study (PEDro scale) only considers 10 categories. This may decrease the quality of the appraisal, however, it does utilize a tool that can provide an objective measure.

Recommendation for Future Research
The volume of research on HIIT has increased significantly over the last 10 years. HIIT has improved the quality of training completed by athletes in many endurance sports. More high quality research is necessary to accurately assess the effects of AeHIIT on endurance performance. Specifically, the quality of studies could be significantly improved by blinding all subjects as well as assessors. Also, all studies should include between-group comparisons. Without comparing groups, it is impossible to determine if one intervention is better than another. Finally, all studies that include two or more exercise programs should use a method to calculate the training load. This would help determine if changes are as a result of the training program or due to different training dosages.

CONCLUSION

·      There is high level evidence (one RCT PEDro 7/10) and moderate level evidence (one RCT PEDro 5/10; one RCT PEDro 4/10) demonstrating that AeHIIT can improve VO2peak in trained male and female cyclists, in state/provincial-level male and female rowers, and endurance-trained male athletes.
·      There is moderate level evidence (one RCT PEDro 510) that demonstrates that AeHIIT can improve stroke volume in endurance-trained male athletes.
·      There is moderate level evidence (one RCT PEDro 5/10) that demonstrates that AeHIIT can improve LT2 (as a percent of VO2peak) in well-trained, competitive endurance athletes (middle distance runners, triathletes and 10 km runners).
·      There is high level evidence (one RCT PEDro 7/10) and moderate level evidence (one RCT PEDro 5/10) that demonstrates that AeHIIT can improve TT performance in well-trained, state/provincial-level male and female rowers, and endurance-trained male athletes, respectively.
·      There is moderate level evidence (two RCTs PEDro 5/10, one RCT PEDro 4/10) that demonstrates that AeHIIT can improve exercise efficiency in trained male and female cyclists, and endurance-trained male athletes.
·      There is moderate level evidence (one RCT PEDro 5/10, one RCT PEDro 4/10) that indicates that AeHIIT improves TTE in trained male and female cyclists, and endurance-trained male athletes.

REFERENCES

1.     Billat L. Interval training for performance: a scientic and empirical practice. Special recommendations for middle- and long-distance running. Part I: Aerobic interval training. Sports Med 2001;31(1):13-31.

2.     Billat L. Interval training for performance: a scientic and empirical practice. Special recommendations for middle- and long-distance running. Part II: Anaerobic interval training. Sports Med 2001;31(2):75-90.

3.     Buchheit M, Laursen P. High-intensity interval training, solutions to the programming puzzle. Part I: Cardiopulmonary emphasis. Sports Med 2013;43(5):313-338.

4.     Buchheit M, Laursen, P. High-intensity interval training, solutions to the programming puzzle. Part II: Anaerobic energy, neuromuscular load and practical applications. Sports Med 2013;43(10):927–954.

5.     Driller M, Fell J, Gregory J, Shing C, Wiliams A. The effects of high-intensity interval training in well trained rowers. Int J Sports Physiol Perform 2009;4:110-121.

6.     Gibala M, Little J, MacDonald M, Hawley J. Physiological adaptations to low-volume, high-intensity interval training in health and disease. J Physiol 2012;590:1077-1084.

7.     Gist N, Fedewa M, Dishman R, Cureton K. Sprint interval training effects on aerobic capacity: a systematic review and meta-analysis. Sports Med 2014;44:269-279.

8.     Helgerud J, Høydal K, Wang E, Karlsen T, Berg P, Bjerkaas M, Simonsen T, Helgesen C, Hjorth N, Bach R, Hoff J. Aerobic high-intensity intervals improve VO2max more than moderate training. Med Sci Sports Exerc 2007;39(4):665-671.

9.     Laursen P, Jenkins D. The scientific basis for high-intensity interval training. Optimising training programmes and maximising performance in highly trained endurance athletes. Sports Med 2002;32(1):53–73.

10.  Seiler S, Jøranson K, Olesen B, Hetlelid K. Adaptations to aerobic interval training: interactive effects of exercise intensity and total work duration. Scand J Med Sci Sports 2013;23(1);74-83.

11.  Smith T, Coombes J, Geraghty D. Optimising high-intensity treadmill training using the running speed at maximal O2 uptake and the time for which this can be maintained. Eur J Appl Physiol 2003;89:337-343.

12.  Stepto N, Hawley J, Dennis S, Hopkins W. Effects of different interval-training programs on cycling time-trial performance. Med Sci Sports Exerc 1999;31(5):736-741.